Managers wish to verify that a particular engineering design meets their require- ments. This design's future environment will differ from the environment assumed during the design. Therefore it is crucial to determine which variations in the envi- ronment may make this design unacceptable. The proposed methodology estimates which uncertain environmental parameters are important (so managers can become pro-active) and which parameter combinations (scenarios) make the design unac- ceptable. The methodology combines simulation, bootstrapping, and metamodeling. The methodology is illustrated through a simulated manufacturing system, includ- ing fourteen uncertain parameters of the input distributions for the various arrival and service times. These parameters are investigated through sixteen scenarios, selected through a two-level fractional-factorial design. The resulting simulation In- put/Output (I/O) data are analyzed through a first-order polynomial metamodel and bootstrapping. A second experiment gives some outputs that are indeed un- acceptable. Polynomials fitted to the I/O data estimate the border line (frontier) between acceptable and unacceptable environments.
|Place of Publication||Tilburg|
|Number of pages||25|
|Publication status||Published - 2009|
|Name||CentER Discussion Paper|
- Uncertainty modeling
- Risk analysis
- Robustness and sensitivity analysis